Increasing believability in animated pedagogical agents

Abstract

Animated pedagogical agents o er great promise for knowledge-based learning environments. In addition to coupling feedback capabilities with a strong visual presence, these agents play a critical role in motivating students. The extent to which they exhibit life-like behaviors strongly increases their motivational impact, but these behaviors must always complement and never interfere with students ' problem solving. To address this problem, we havedeveloped a framework for dynamically sequencing animated pedagogical agents ' believability-enhancing behaviors. By monitoring a student's problemsolving history and the agent's past activities, a competition-based behavior sequencing engine produces realtime life-like character animations that are pedagogically appropriate. Behaviors in the agent's repertoire compete with one another. At each moment, the strongest eligible behavior is heuristically selected as the winner and is exhibited. We haveimplemented this framework in Herman the Bug, an animated pedagogical agent that inhabits a knowledge-based learning environment for the domain of botanical anatomy and physiology.